Treating Intraoperative Hyperchloremic Acidosis with Sodium ... : Anesthesia & Analgesia (original) (raw)
We demonstrated that saline infusion of approximately 35 mL/kg within 2 h during anesthesia and surgery will inevitably lead to hyperchloremic acidosis in a dose-dependent manner (1). Also, the infusion of saline-based colloid solutions will result in hyperchloremic acidosis (2). In this context, the seemingly harmless and benign character of this acid-base disturbance was called into question (3–5). Still, severe hyperchloremic acidosis should be avoided. One possibility to reach this target is the use of balanced solutions, such as lactated Ringer’s solution or Hextend®(1,3,4). The second possibility is correcting this acidosis at an early state. Until now, however, there was an open discussion as to the better concept for treating this kind of acidosis: with sodium bicarbonate (BIC) or tris-hydroxymethyl aminomethane (THAM) (6). Undoubtedly, both drugs may lead to a correction of the acidosis; however, the respective effects on mechanical ventilation aiming at constant Paco2 and serum concentrations of electrolytes or unmeasured ions should be very different. By comparing two groups of patients with intraoperative hyperchloremic acidosis caused by 0.9% saline infusion and randomly assigned to receive equal doses of either BIC or THAM, we assessed the effect of both drugs on all important variables of acid-base chemistry at 0, 10, and 20 min after buffering. For evaluating the causes of the changes in pH after buffering, we planned to interpret the changes in serum bicarbonate concentration [Bic] and base excess [BE] as calculated by the Henderson-Hasselbalch equation and the Siggaard-Andersen nomogram (7) in the light of the Stewart (8) approach to acid-base chemistry. For more clarity of the following passage, all abbreviations (and calculations) are explained in Table 1. According to Stewart’s paradigms, Paco2, the sum of all anionic charges of weak plasma acids [A−] (called Atot by Stewart), and the strong ion difference (SID) were defined as independent pH-regulating variables, whereas pH and [Bic] were only dependent variables. In Stewart’s terminology, weak plasma acids are incompletely dissociated substances in blood plasma (such as albumin or phosphate), whereas strong ions are completely dissociated electrolytes and lactate. Stewart calculated [A−] as the product of total protein and the empirically derived factor 2.43 according to van Slyke et al. (9). Figge et al. (10), however, developed a special computer program considering the net negative charges on albumin and phosphate, which are the main weak negative charges in blood plasma. Stewart’s approach postulates that a decrease in pH is the result of a decrease in SID or an increase in [A−] on the condition that Paco2 is constant (8). Figure 1 shows an ionogram of an exemplary hyperchloremic acidosis. The basis of this ionogram is the electroneutrality in blood plasma, so that the sum of positive charges equals the sum of negative charges. As shown in Figure 1, there are two possibilities to quantify SID. One is to calculate the difference of the measured strong ions: serum sodium [Na+] plus serum potassium [K+] minus serum chloride [Cl−] minus serum lactate [Lac−] (see also Table 1). This can be done at the bedside and gives an apparent SID (SID a). The second way is to use the computer program by Figge et al. (10), which calculates an effective SID (SID e) as a function of pH, Paco2, plasma albumin, and phosphate concentration according to the Stewart approach. Normally, SID a is larger than SID e because of the existence of some unidentified negative charges (unmeasured strong anions) in blood plasma. As may be taken from Figure 1, the difference between SID a and SID e, the strong ion gap (SIG), can be used to quantify unidentified strong anions (11,12). According to Stewart’s algorithms, the mechanism behind acidosis after 0.9% saline infusion is a decrease in SID (SID a and SID e) caused by hyperchloremia (1). The main objective of this study was to elucidate the mechanisms behind the changes in pH after buffering with BIC or THAM by using the Stewart approach. According to Stewart’s postulate, an increase in pH should be the result of an increase in SID or a decrease in [A−] on condition of constant Paco2(8). This is the first investigation that evaluates the effects of buffer solutions in the context of Stewart’s paradigms.
Abbreviations and Calculations
Ionogram of an exemplary hyperchloremic acidosis caused by infusion of 0.9% saline. [Bic] = serum bicarbonate concentration (dependent variable according to Stewart) (8); [A−] = sum of all anionic charges of weak plasma acids; SID a = apparent (bedside) strong ion difference, calculation: SID a = [Na+] + [K+] − [Cl−] − [Lac−]; SID e = effective strong ion difference, according to the computer program of Figge et al. (10) ([A−] and SID are independent variables according to Stewart) (8); strong ion gap = SID a − SID e; [Na+] = serum concentration of sodium; [Cl−] = serum concentration of chloride; [K+] = serum concentration of potassium; [Lac−] = serum concentration of lactate.
Methods
We studied 24 female patients who had cancer of the ovary without apparent cardiac, pulmonary, or renal diseases (ASA physical status II) and who were scheduled for radical or cytoreductive gynecological abdominal surgery. All patients had normal preoperative acid-base and electrolyte values. Written, informed consent was obtained from each patient before surgery, and the protocol was approved by the ethics committee of our institution. During the study, no patient received colloids, plasma products, or blood transfusions.
After the induction of general anesthesia and endotracheal intubation, radial arterial and central venous catheters were inserted. Mechanical ventilation was performed with 100% oxygen, and we intended to continuously maintain Paco2 as close as possible to 40 mm Hg (for abbreviations, see Table 1). During the operative period, the patient’s temperature was kept constant with fluid warmers and warming blankets.
At the time of surgical incision, the patients received, for 2 h, 40 mL · kg−1 · h−1 of 0.9% saline solution containing 154 mmol/L of sodium and 154 mmol/L of chloride. This large volume of crystalloid infusion seemed appropriate because patients with ovarian cancer need especially large intraoperative infusion volumes to maintain hemodynamic stability and normovolemia (13) and because we did not want to use any colloid in this early operative period. In every patient, hyperchloremic acidosis occurred, and, at a standard BE of −7 mmol/L, arterial blood samples were taken to obtain baseline values.
The samples were analyzed for Pao2, pH, Paco2 (standard electrodes), [Na+], [Cl−], [K+], ionized calcium [Ca2+] (ion-elective electrodes), and [Lac−] (enzymatic method, quantification of H2O2), all integrated in a Radiometer analyzer (Radiometer ABL 620 GL; Radiometer Co., Copenhagen, Denmark). Additionally, serum phosphate [PO4] (ultraviolet photometry of a phosphomolybdate complex), serum total protein concentration (Biuret method), albumin concentration (colorimetry of bromocresol complex), and serum glucose concentration (hexokinase method) were measured from the same blood samples. [Bic] and [BE] were taken from the blood-gas analyzer, which uses the Henderson-Hasselbalch equation and the Siggaard-Andersen nomogram (10). For each sample, SID a, SID e, SIG, and anion gap were calculated (for abbreviations and calculations, see Table 1).
After the baseline blood samples had been taken, the patients were randomly assigned to receive either BIC 1 M (Braun Melsungen AG, Melsungen, Germany) (BIC group;n = 12) or THAM 3 M (Braun Melsungen AG) (THAM group;n = 12). In both groups, the doses of BIC or THAM infused were calculated as
The extracellular volume (ECV) was calculated as
Both drugs were infused through the central venous catheter within exactly 20 min. In addition to the infusion of the drugs, we planned to infuse 0.9% saline through the same line to decrease the high osmolarity of the buffer solutions to approximately 600 mM. In the BIC group, the volume of saline should be approximately four times the volume of 1 M BIC, and in the THAM group, the volume of saline should be approximately nine times the volume of 3 M THAM. By adapting the mechanical ventilation closely to the end-tidal Pco2, we also tried to maintain Paco2 as close as possible to 40 mm Hg during and after the infusion of BIC or THAM. After the infusion of BIC or THAM, the remaining flow of crystalloid infusion was 5 mL/min. At 0, 10, and 20 min after the infusion of BIC or THAM, blood samples were taken again, and all variables mentioned above were measured in the same way as for the baseline measurements.
Because all measured and calculated data described above were normally distributed (tested by Kolmogorov-Smirnov tests), they are presented as mean and sd. For demographic data, Student’s _t_-tests for unpaired data were performed. Two-way repeated-measures analysis of variance compared intragroup and intergroup differences at baseline and at 0, 10, and 20 min after BIC or THAM administration. Post hoc testing was performed according to the Student-Newman-Keuls method for multiple comparisons. P < 0.05 was considered significant.
Results
There were no significant differences between groups with respect to demographic data (mean values ± sd; BIC group: age, 55 ± 15 yr; height, 165 ± 7 cm; weight, 66 ± 7 kg; THAM group: age, 54 ± 18 yr; height, 164 ± 7 cm; weight, 66 ± 11 kg). Table 2 shows the amount of 0.9% saline infused, urine production, estimated blood loss, amounts of BIC and THAM infused, minute ventilation (MV), end-tidal Pco2, and arterial oxygen partial pressure (Pao2) at the various measuring points. At the baseline measurements, the amount of saline infused (within 2 h) was similar in both groups, i.e., approximately 5.200 mL. To reduce the osmolarity of the buffer solutions infused, approximately 400 or 500 mL of 0.9% saline was administered during the infusion of BIC or THAM. From 0 to 20 min, in both groups only minor volumes of crystalloid were infused. In both groups, baseline values for urine production and estimated blood loss amounted to approximately 600 and 550 mL, respectively, and increased until 20 min after the infusion of BIC or THAM to approximately 900 and 750 mL, respectively. In no patient did hemodynamic data (not presented) show any signs of hyper- or hypovolemia at any measuring point. During the infusion of a mean of 128 ± 18 mmol of BIC, MV had to be increased by approximately 40% to maintain a constant end-tidal Pco2. However, 10 min after the infusion of BIC, MV reached the baseline value. During the infusion of 130 ± 26 mmol of THAM, MV had to be decreased by approximately 60%, and 10 min later, near-baseline MV could be used. The only significant intergroup difference was found in MV at 0 min after the buffer infusion. Neither the infusion of BIC nor that of THAM, involving the inverse changes in MV, resulted in a significant decrease in Pao2.
Volume Balance, Amount of Buffer Solutions Infused, and Variables Concerning Mechanical Ventilation (Mean ± sd)
Table 3 shows pH, Paco2, [Bic], [BE], [Na+], [Cl−], [K+], [Lac−], SID a, SID e, SIG, [A−], [PO4], [Ca2+], anion gap, and serum glucose concentration at the different measuring points separately for both groups. At the baseline measuring point, no significant differences between groups could be found, except that baseline [Na+] was slightly higher in the BIC group (+3 mmol/L) than in the THAM group. A retrospective analysis showed that there was already a small preoperative difference in [Na+] before the saline infusion was started.
Measured and Calculated Values of Acid-Base State (Mean ± sd)
In the BIC group, because of the infusion of the buffer solution, pH increased from a mean of 7.278 to 7.440, and Paco2 remained fairly constant at 40 mm Hg. Mean [Bic] and [BE] increased by 8.1 and 9.8 mmol/L, respectively. The increasing [Na+] (+5 mmol/L) and decreasing [Cl−] (−3 mmol/L) at almost constant [K+] and [Lac−] resulted in a significant increase in SID a (+7.8 mEq/L) 0 min after the infusion of BIC. The mean increase in SID e, amounting to +8.6 mEq/L, was absolutely comparable to the increase in SID a. SIG, [A−], [PO4], [Ca2+], anion gap, and serum glucose concentration did not change significantly after the infusion of BIC. The comparison of the values at 10 and 20 min with the values at 0 min (in the BIC group) shows a slight but continuous decrease in pH, [Bic], [BE], SID a, and SID e.
In the THAM group, 0 min after the infusion of the buffer solution, pH had increased from a mean of 7.280 to 7.400, whereas Paco2 remained fairly constant at 40 mm Hg during the study. [Bic] increased by a mean of 5.9 mmol/L and [BE] by 7.2 mmol/L. From baseline to 0 min after the infusion, [Na+] and [Cl−] decreased by 2 mmol/L each, and because of constant [K+] and [Lac−], this resulted in an almost unchanged SID a (+0.3 mEq/L). In contrast, SID e increased significantly by a mean of +6.4 mEq/L after the infusion of THAM. A constant SID a and increased SID e resulted in a significant decrease in SIG (−6.2 mEq/L). This decrease in SIG, as well as the decrease in the anion gap (−5.8 mmol/L), can be explained by the occurrence of an unmeasured cation after the infusion of THAM. [A−], [PO4], [Ca2+], and serum glucose concentration did not change significantly after the infusion of THAM. In the THAM group, the comparison of the values at 10 and 20 min with the values at 0 min shows a slight but continuous decrease in pH, [Bic], and [BE] in a comparable amount as in the BIC group. SID a remained almost constant in this period after the infusion of THAM, whereas SID e slightly but continuously decreased and SIG, as well as the anion gap, increased.
Discussion
The infusion of BIC or THAM resulted in a complete correction of metabolic acidosis. With constant Paco2 and unchanged values for [A−], the increase in pH in both groups was associated with an increase in SID (for abbreviations and calculations, see Table 1). After the infusion of BIC, SID a and SID e increased, whereas after the infusion of THAM—because of the occurrence of an unmeasured cation—SID e increased at constant SID a. In both groups, however, hyperchloremia was not corrected.
Because equal doses of BIC or THAM were administered in both groups, a dose-effect comparison of both drugs concerning almost all important variables of acid-base equilibrium can be drawn. The increase in pH, [Bic], and [BE] was slightly larger after the administration of BIC in comparison with THAM. The reason for this difference can be found in the incomplete dissociation of THAM (only approximately 70%) (14). One of the most impressive differences between the drugs is the different reaction with carbon dioxide. BIC infusion results in an increase in carbon dioxide production according to Equation 1:
In contrast, the infusion of THAM reduces pulmonary carbon dioxide output, as may be taken from Equation 2(14):
where R-NH2 indicates unprotonated THAM and R-NH3+ indicates protonated THAM.
As a consequence, mechanical ventilation had to be increased during the administration of BIC (approximately +40%) and decreased during the infusion of THAM (approximately −60%) to prevent any change in Paco2. However, in both groups, the effects on CO2 were only transient.
Because Paco2 was continuously kept constant in both groups, the observed acid-base changes after BIC or THAM infusion were undoubtedly of metabolic origin. Concerning the question of the mechanisms behind the metabolic changes in acid-base chemistry after buffering with BIC or THAM, the conventional Siggaard-Andersen approach (7) would tell us that only a supply of bicarbonate could be the reason for the observed increase in pH and [Bic] in both groups. In case of BIC infusion, there was a direct supply of bicarbonate (see Equation 1). The infusion of THAM should result in an indirect bicarbonate supply in the form of “bicarbonate production” according to the Equation 2. However, there is an alternative to explain the observed metabolic changes in acid-base chemistry: Stewart’s approach (8), which was discussed in detail elsewhere (16), defines Paco2, SID, and the sum of all anionic charges of [A−] as independent pH-regulating variables, whereas pH and [Bic] are dependent variables. It was uncommon to apply Stewart’s approach until 1999. Recently, however, this approach has attracted considerable attention, partly because of the promise of explaining, in a relatively simple quantitative way, the whole mystery of acid-base chemistry (17–22).
Stewart’s approach postulates that a decrease in pH (on the condition that Paco2 is constant) is the result of an increase in SID or a decrease in [A−] (and vice versa) (8). In a study from our laboratory, we have shown the dose-dependent occurrence of metabolic acidosis after rapid intraoperative saline infusion (1). This acidosis was connected with hyperchloremia, a profound decrease in bedside or SID a, and a concomitant but smaller decrease in total weak plasma acid, termed Prot− now termed [A−] (Fig. 1). After the administration of lactated Ringer’s solution, both SID a and [A−] decreased by the same amount, so that no change in Bic occurred (1). In a following study (2) using different colloid solutions, one containing albumin in a supraphysiological concentration and one free of albumin, reverse changes in [A−] and, because both were saline-based solutions, concomitant changes in SID could be observed. In this study, however, the almost unique situation of nearly constant Paco2 and constant [A−] could be investigated during large changes in pH and [Bic] after the infusion of BIC or THAM. Therefore, according to the Stewart approach (8), only changes in SID should explain the changes in pH and [Bic] in this specific study design. In other words, by defining [Bic] as being exclusively dependent on SID and [A−] during metabolic acid-base changes, any supply of bicarbonate is completely irrelevant according to Stewart’s paradigms.
Figure 2 illustrates separately for both groups the changes in [Bic] in the context of the changes in SID a and SID e at the different measuring points. In the BIC group, the changes in [Bic] were always in good agreement with the respective changes in SID a and SID e. In contrast, in the case of THAM infusion, the bedside SID a did not reflect the changes in acid-base balance, whereas the changes in SID e were in good agreement with the observed changes in [Bic]. Consequently, the Stewart approach will fail in case of THAM infusion if SID a is used to explore the cause for the respective change in acid-base equilibrium. After the infusion of THAM, the difference between SID a and SID e, i.e., SIG, decreased significantly. The reason for the phenomenon of a change in SIG should be the occurrence of unmeasured ions after the infusion of THAM. Often, the anion gap (for calculation, see Table 1) is used to explore unexplained ions; however, SIG can also be useful for this purpose (11,12). Figure 3 shows changes in anion gap and changes in SIG in the THAM group. It can be seen that the decreases in anion gap and SIG were almost identical. Fortunately, in the special case of THAM infusion, this unmeasured ion can be explained and identified. It should be simply the protonized THAM+ molecule.
Mean changes in serum bicarbonate [Bic], apparent strong ion difference (SID a), and effective strong ion difference (SID e) due to the infusion of sodium bicarbonate (BIC) or tris-hydroxymethyl aminomethane (THAM).
Mean changes in anion gap ([Na+] + [K+] − [Cl−] − [Bic]) and strong ion gap (SIG; SID a − SID e) due to the infusion of tris-hydroxymethyl aminomethane (THAM). [Na+] = serum concentration of sodium; [Cl−] = serum concentration of chloride; [K+] = serum concentration of potassium; [Bic] = serum bicarbonate concentration; SID a = apparent (bedside) strong ion difference; SID e = effective strong ion difference.
As a result, in case of buffering with BIC or THAM, the respective changes in acid-base equilibrium were accompanied by an increase in SID. In the special situation of THAM infusion, however, only the change in SID e could reflect the respective changes in [Bic]. In the case of BIC administration, the infusion of a strong ion in the form of sodium, and, in the case of THAM administration, the infusion of an unmeasured cation, was the reason for the increase in SID e and, according to the Stewart terminology, in [Bic]. This is the first description of an unmeasured cation (in the context of THAM infusion) with the Stewart approach.
One pragmatic question after this study is which one is the better choice for treating intraoperative hyperchloremic acidosis caused by saline infusion: BIC or THAM? First, it must be stated that this study can offer only a small step toward the difficult answer to this possibly important question. In clinical practice, BIC seems to be used more often than THAM for treating intraoperative hyperchloremic acidosis. By means of both drugs, our patients’ acidosis could be corrected; however, neither the infusion of BIC nor that of THAM corrected hyperchloremia significantly. BIC seems to be more effective, but it can lead to hypernatremia. During the infusion of BIC, mechanical ventilation had to be increased significantly to keep Paco2 constant. This involves some limitations in patients with severe pulmonary disease. As already demonstrated in a rabbit model (23), we could not observe hypokalemia or even hypoglycemia in our patients after BIC or THAM infusion.
There is still controversy about the place of THAM in the management of acidemia (24,25). THAM could offer some advantages, because it does not lead to hypernatremia and because pulmonary carbon dioxide output decreases. As a consequence, THAM seems to be a good alternative, especially in patients with respiratory failure and those in whom hypernatremia is a concern.
The last important question is which possibility should be preferred: avoiding hyperchloremic acidosis by using balanced infusion solutions (3) or — as in our study — correcting it early? Unfortunately, we cannot answer this question with our data. Probably, avoiding intraoperative hyperchloremic acidosis should be preferred; however, until now, there was no clear evidence for a better postoperative outcome of this method (3–5). As a result, intense systematic analyses of this topic are necessary.
We would like to thank Dr. Fencl for allowing us to work with his computer program.
References
1. Scheingraber S, Rehm M, Sehmisch C, Finsterer U. Rapid saline infusion produces hyperchloremic acidosis in patients undergoing gynecologic surgery. Anesthesiology 1999; 90: 1265–70.
2. Rehm M, Orth V, Scheingraber S, et al. Acid-base changes caused by 5% albumin versus 6% hydroxyethylstarch solution in patients undergoing acute normovolemic hemodilution. Anesthesiology 2000; 93: 1174–83.
3. Wilkes NJ, Wolf R, Mutch M, et al. The effects of balanced versus saline-based hetastarch and crystalloid solutions on acid-base and electrolyte status and gastric mucosal perfusion in elderly surgical patients. Anesth Analg 2001; 93: 811–6.
4. Waters J, Gottlieb A, Schoenwald P, et al. Normal saline versus lactated Ringer’s solution for intraoperative fluid management in patients undergoing abdominal aortic aneurysm repair: an outcome study. Anesth Analg 2001; 93: 817–22.
5. O’Connor MF, Roizen MF. Lactate versus chloride: which is better [editorial]? Anesth Analg 2001; 93: 809–10.
6. Roth VR. What is the clinical relevance of dilutional acidosis [letter; replies]? Anesthesiology 2001; 95: 810–2.
7. Siggaard-Andersen O. The acid-base status of the blood. 4th ed. Copenhagen: Munksgaard, 1976.
8. Stewart PA. Modern quantitative acid-base chemistry. Can J Physiol Pharmacol 1983; 61: 1444–61.
9. van Slyke DD, Hastings AB, Hiller A, Sendroy J. Studies of gas and electrolyte equilibria in blood. XIV. Amount of alkali bound by serum albumin and globulin. J Biol Chem 1928; 79: 769–80.
10. Figge J, Mydosh T, Fencl V. Serum proteins and acid-base equilibria: a follow-up. J Lab Clin Med 1992; 120: 713–9.
11. Gilfix BM, Bique M, Magder S. A physical chemical approach to the analysis of acid-base balance in the clinical setting. J Crit Care 1993; 8: 187–97.
12. Kellum JA, Kramer DJ, Pinsky MR. Strong ion gap: a methodology for exploring unexplained anions. J Crit Care 1995; 10: 51–5.
13. Rehm M, Haller M, Brechtelsbauer H, et al. Extra protein loss not caused by surgical bleeding in patients with ovarian cancer. Acta Anaesthesiol Scand 1998; 42: 39–46.
14. Nahas GG, Sutin KM, Fermon C, et al. Guidelines for the treatment of acidemia with THAM. Drugs 1998; 55: 191–224.
15. Madias NE, Cohen JJ. Acid-base chemistry and buffering. In: Cohen JJ, Kassirer JP, eds. Acid-base. Boston: Little, Brown, 1982: 3–25.
16. Fencl V, Leith DE. Stewart’s quantitative acid-base chemistry: applications in biology and medicine. Respir Physiol 1993; 91: 1–16.
17. Kellum JA. Acid-base physiology in the post-Copernican era. Curr Opin Crit Care 1999; 5: 429–35.
18. Liskaser FJ, Bellomo R, Hayhoe M, et al. The role of pump prime in the etiology and pathogenesis of cardiopulmonary bypass-associated acidosis. Anesthesiology 2000; 93: 1170–3.
19. Waters JH, Bernstein CA. Dilutional acidosis following hetastarch or albumin in healthy volunteers. Anesthesiology 2000; 93: 1184–7.
20. Fencl V, Jabor A, Kazda A, Figge J. Diagnosis of metabolic acid-base disturbances in critically ill patients. Am J Respir Crit Care Med 2000; 162: 2246–51.
21. Morgan TJ, Balasubramanian V, Hall J. Crystalloid strong ion difference determines metabolic acid-base change during in vitro hemodilution. Crit Care Med 2002; 30: 157–60.
22. Kellum JA. Saline-induced hyperchloremic metabolic acidosis [editorial]. Crit Care Med 2002; 30: 259–61.
23. Waters JH, Howard RS, Lesnik IK. Plasma potassium response after tromethamine (THAM) or sodium bicarbonate in the acidotic rabbit. Anesth Analg 1996; 83: 789–92.
24. Adrogué AJ, Madias NE. Management of life-threatening acid-base disorders. N Engl J Med 1998; 338: 26–34.
25. Holmdahl MH, Wiklund L, Wetterberg T, et al. The place of THAM in the management of acidemia in clinical practice. Acta Anaesthesiol Scand 2000; 44: 524–7.
© 2003 International Anesthesia Research Society